refactor: cleanup code

du has -L flag which allows it to dereference or follow
symlinks. Clone this feature into dust.
https://github.com/bootandy/dust/issues/276

Cargo.lock

@@ -118,6 +118,12 @@ dependencies = [
  "toml",
 ]
 
+[[package]]
+name = "core-foundation-sys"
+version = "0.8.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5827cebf4670468b8772dd191856768aedcb1b0278a04f989f7766351917b9dc"
+
 [[package]]
 name = "crossbeam-channel"
 version = "0.5.6"
@@ -197,10 +203,11 @@ checksum = "fea41bba32d969b513997752735605054bc0dfa92b4c56bf1189f2e174be7a10"
 
 [[package]]
 name = "du-dust"
-version = "0.8.2"
+version = "0.8.3"
 dependencies = [
  "ansi_term",
  "assert_cmd",
+ "atty",
  "clap",
  "clap_complete",
  "config-file",
@@ -210,6 +217,7 @@ dependencies = [
  "regex",
  "serde",
  "stfu8",
+ "sysinfo",
  "tempfile",
  "terminal_size",
  "thousands",
@@ -322,6 +330,15 @@ dependencies = [
  "autocfg",
 ]
 
+[[package]]
+name = "ntapi"
+version = "0.4.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "bc51db7b362b205941f71232e56c625156eb9a929f8cf74a428fd5bc094a4afc"
+dependencies = [
+ "winapi",
+]
+
 [[package]]
 name = "num_cpus"
 version = "1.13.1"
@@ -518,6 +535,21 @@ dependencies = [
  "unicode-ident",
 ]
 
+[[package]]
+name = "sysinfo"
+version = "0.26.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c375d5fd899e32847b8566e10598d6e9f1d9b55ec6de3cdf9e7da4bdc51371bc"
+dependencies = [
+ "cfg-if",
+ "core-foundation-sys",
+ "libc",
+ "ntapi",
+ "once_cell",
+ "rayon",
+ "winapi",
+]
+
 [[package]]
 name = "tempfile"
 version = "3.3.0"

Cargo.toml

@@ -1,7 +1,7 @@
 [package]
 name = "du-dust"
 description = "A more intuitive version of du"
-version = "0.8.2"
+version = "0.8.3"
 authors = ["bootandy <bootandy@gmail.com>", "nebkor <code@ardent.nebcorp.com>"]
 edition = "2021"
 readme = "README.md"
@@ -28,6 +28,7 @@ strip = true
 
 [dependencies]
 ansi_term = "0.12"
+atty = "0.2.14"
 clap = "3.2.17"
 lscolors = "0.7"
 terminal_size = "0.1"
@@ -39,6 +40,7 @@ regex = "1"
 config-file = "0.2"
 serde = { version = "1.0", features = ["derive"] }
 directories = "4"
+sysinfo = "0.26"
 
 [target.'cfg(windows)'.dependencies]
 winapi-util = "0.1"

README.md

@@ -30,6 +30,10 @@ Because I want an easy way to see where my disk is being used.
 
 - `pacstall -I dust-bin`
 
+#### [deb-get](https://github.com/wimpysworld/deb-get) (Debian/Ubuntu)
+
+- `deb-get install du-dust`
+
 #### Windows:
 
 - Windows GNU version - works
@@ -59,6 +63,7 @@ Usage: dust -p (full-path - Show fullpath of the subdirectories)
 Usage: dust -s (apparent-size - shows the length of the file as opposed to the amount of disk space it uses)
 Usage: dust -n 30  (Shows 30 directories instead of the default [default is terminal height])
 Usage: dust -d 3  (Shows 3 levels of subdirectories)
+Usage: dust -D (Show only directories (eg dust -D))
 Usage: dust -r (reverse order of output)
 Usage: dust -H (si print sizes in powers of 1000 instead of 1024)
 Usage: dust -X ignore  (ignore all files and directories with the name 'ignore')

completions/_dust

@@ -35,6 +35,10 @@ _dust() {
 '--version[Print version information]' \
 '-p[Subdirectories will not have their path shortened]' \
 '--full-paths[Subdirectories will not have their path shortened]' \
+'(-L --dereference-links)-l[Ignore links]' \
+'(-L --dereference-links)--ignore-links[Ignore links]' \
+'(-l --ignore-links)-L[dereference sym links - Treat sym links as directories and go into them]' \
+'(-l --ignore-links)--dereference-links[dereference sym links - Treat sym links as directories and go into them]' \
 '-x[Only count the files and directories on the same filesystem as the supplied directory]' \
 '--limit-filesystem[Only count the files and directories on the same filesystem as the supplied directory]' \
 '-s[Use file length instead of blocks]' \
@@ -54,6 +58,8 @@ _dust() {
 '(-d --depth)--file_types[show only these file types]' \
 '-H[print sizes in powers of 1000 (e.g., 1.1G)]' \
 '--si[print sizes in powers of 1000 (e.g., 1.1G)]' \
+'-D[Only directories will be displayed.]' \
+'--only-dir[Only directories will be displayed.]' \
 '*::inputs:' \
 && ret=0
 }

completions/_dust.ps1

@@ -41,6 +41,10 @@ Register-ArgumentCompleter -Native -CommandName 'dust' -ScriptBlock {
             [CompletionResult]::new('--version', 'version', [CompletionResultType]::ParameterName, 'Print version information')
             [CompletionResult]::new('-p', 'p', [CompletionResultType]::ParameterName, 'Subdirectories will not have their path shortened')
             [CompletionResult]::new('--full-paths', 'full-paths', [CompletionResultType]::ParameterName, 'Subdirectories will not have their path shortened')
+            [CompletionResult]::new('-l', 'l', [CompletionResultType]::ParameterName, 'Ignore links')
+            [CompletionResult]::new('--ignore-links', 'ignore-links', [CompletionResultType]::ParameterName, 'Ignore links')
+            [CompletionResult]::new('-L', 'L', [CompletionResultType]::ParameterName, 'dereference sym links - Treat sym links as directories and go into them')
+            [CompletionResult]::new('--dereference-links', 'dereference-links', [CompletionResultType]::ParameterName, 'dereference sym links - Treat sym links as directories and go into them')
             [CompletionResult]::new('-x', 'x', [CompletionResultType]::ParameterName, 'Only count the files and directories on the same filesystem as the supplied directory')
             [CompletionResult]::new('--limit-filesystem', 'limit-filesystem', [CompletionResultType]::ParameterName, 'Only count the files and directories on the same filesystem as the supplied directory')
             [CompletionResult]::new('-s', 's', [CompletionResultType]::ParameterName, 'Use file length instead of blocks')
@@ -60,6 +64,8 @@ Register-ArgumentCompleter -Native -CommandName 'dust' -ScriptBlock {
             [CompletionResult]::new('--file_types', 'file_types', [CompletionResultType]::ParameterName, 'show only these file types')
             [CompletionResult]::new('-H', 'H', [CompletionResultType]::ParameterName, 'print sizes in powers of 1000 (e.g., 1.1G)')
             [CompletionResult]::new('--si', 'si', [CompletionResultType]::ParameterName, 'print sizes in powers of 1000 (e.g., 1.1G)')
+            [CompletionResult]::new('-D', 'D', [CompletionResultType]::ParameterName, 'Only directories will be displayed.')
+            [CompletionResult]::new('--only-dir', 'only-dir', [CompletionResultType]::ParameterName, 'Only directories will be displayed.')
             break
         }
     })

completions/dust.bash

@@ -19,7 +19,7 @@ _dust() {
 
     case "${cmd}" in
         dust)
-            opts="-h -V -d -n -p -X -x -s -r -c -b -z -f -i -v -e -t -w -H --help --version --depth --number-of-lines --full-paths --ignore-directory --limit-filesystem --apparent-size --reverse --no-colors --no-percent-bars --min-size --skip-total --filecount --ignore_hidden --invert-filter --filter --file_types --terminal_width --si <inputs>..."
+            opts="-h -V -d -n -p -X -l -L -x -s -r -c -b -z -f -i -v -e -t -w -H -D --help --version --depth --number-of-lines --full-paths --ignore-directory --ignore-links --dereference-links --limit-filesystem --apparent-size --reverse --no-colors --no-percent-bars --min-size --skip-total --filecount --ignore_hidden --invert-filter --filter --file_types --terminal_width --si --only-dir <inputs>..."
             if [[ ${cur} == -* || ${COMP_CWORD} -eq 1 ]] ; then
                 COMPREPLY=( $(compgen -W "${opts}" -- "${cur}") )
                 return 0

completions/dust.elv

@@ -38,6 +38,10 @@ set edit:completion:arg-completer[dust] = {|@words|
             cand --version 'Print version information'
             cand -p 'Subdirectories will not have their path shortened'
             cand --full-paths 'Subdirectories will not have their path shortened'
+            cand -l 'Ignore links'
+            cand --ignore-links 'Ignore links'
+            cand -L 'dereference sym links - Treat sym links as directories and go into them'
+            cand --dereference-links 'dereference sym links - Treat sym links as directories and go into them'
             cand -x 'Only count the files and directories on the same filesystem as the supplied directory'
             cand --limit-filesystem 'Only count the files and directories on the same filesystem as the supplied directory'
             cand -s 'Use file length instead of blocks'
@@ -57,6 +61,8 @@ set edit:completion:arg-completer[dust] = {|@words|
             cand --file_types 'show only these file types'
             cand -H 'print sizes in powers of 1000 (e.g., 1.1G)'
             cand --si 'print sizes in powers of 1000 (e.g., 1.1G)'
+            cand -D 'Only directories will be displayed.'
+            cand --only-dir 'Only directories will be displayed.'
         }
     ]
     $completions[$command]

completions/dust.fish

@@ -8,6 +8,8 @@ complete -c dust -s w -l terminal_width -d 'Specify width of output overriding t
 complete -c dust -s h -l help -d 'Print help information'
 complete -c dust -s V -l version -d 'Print version information'
 complete -c dust -s p -l full-paths -d 'Subdirectories will not have their path shortened'
+complete -c dust -s l -l ignore-links -d 'Ignore links'
+complete -c dust -s L -l dereference-links -d 'dereference sym links - Treat sym links as directories and go into them'
 complete -c dust -s x -l limit-filesystem -d 'Only count the files and directories on the same filesystem as the supplied directory'
 complete -c dust -s s -l apparent-size -d 'Use file length instead of blocks'
 complete -c dust -s r -l reverse -d 'Print tree upside down (biggest highest)'
@@ -18,3 +20,4 @@ complete -c dust -s f -l filecount -d 'Directory \'size\' is number of child fil
 complete -c dust -s i -l ignore_hidden -d 'Do not display hidden files'
 complete -c dust -s t -l file_types -d 'show only these file types'
 complete -c dust -s H -l si -d 'print sizes in powers of 1000 (e.g., 1.1G)'
+complete -c dust -s D -l only-dir -d 'Only directories will be displayed.'

src/cli.rs

@@ -34,6 +34,20 @@ pub fn build_cli() -> Command<'static> {
                 .multiple_occurrences(true)
                 .help("Exclude any file or directory with this name"),
         )
+        .arg(
+            Arg::new("ignore_links")
+                .short('l')
+                .long("ignore-links")
+                .conflicts_with("dereference_links")
+                .help("Ignore links"),
+        )
+        .arg(
+            Arg::new("dereference_links")
+                .short('L')
+                .long("dereference-links")
+                .conflicts_with("ignore_links")
+                .help("dereference sym links - Treat sym links as directories and go into them"),
+        )
         .arg(
             Arg::new("limit_filesystem")
                 .short('x')
@@ -131,5 +145,11 @@ pub fn build_cli() -> Command<'static> {
                 .long("si")
                 .help("print sizes in powers of 1000 (e.g., 1.1G)")
         )
-        .arg(Arg::new("inputs").multiple_occurrences(true).default_value("."))
+        .arg(Arg::new("inputs").multiple_occurrences(true))
+        .arg(
+            Arg::new("only_dir")
+                .short('D')
+                .long("only-dir")
+                .help("Only directories will be displayed."),
+        )
 }

src/config.rs

@@ -19,6 +19,7 @@ pub struct Config {
     pub ignore_hidden: Option<bool>,
     pub iso: Option<bool>,
     pub min_size: Option<String>,
+    pub only_dir: Option<bool>,
 }
 
 impl Config {
@@ -61,6 +62,9 @@ impl Config {
             size_from_param
         }
     }
+    pub fn get_only_dir(&self, options: &ArgMatches) -> bool {
+        Some(true) == self.only_dir || options.is_present("only_dir")
+    }
 }
 
 fn convert_min_size(input: &str, iso: bool) -> Option<usize> {

src/dir_walker.rs

@@ -26,6 +26,8 @@ pub struct WalkData<'a> {
     pub use_apparent_size: bool,
     pub by_filecount: bool,
     pub ignore_hidden: bool,
+    pub ignore_links: bool,
+    pub follow_links: bool,
 }
 
 pub fn walk_it(dirs: HashSet<PathBuf>, walk_data: WalkData) -> (Vec<Node>, bool) {
@@ -141,24 +143,27 @@ fn walk(
 
                     // return walk(entry.path(), permissions_flag, ignore_directories, allowed_filesystems, use_apparent_size, by_filecount, ignore_hidden);
 
-                    if !ignore_file(entry, walk_data) {
-                        if let Ok(data) = entry.file_type() {
-                            return if data.is_dir() && !data.is_symlink() {
-                                walk(entry.path(), permissions_flag, walk_data, depth + 1)
-                            } else {
-                                build_node(
-                                    entry.path(),
-                                    vec![],
-                                    walk_data.filter_regex,
-                                    walk_data.invert_filter_regex,
-                                    walk_data.use_apparent_size,
-                                    data.is_symlink(),
-                                    data.is_file(),
-                                    walk_data.by_filecount,
-                                    depth,
-                                )
-                            };
+                    if ignore_file(entry, walk_data) {
+                        return None;
+                    }
+                    if let Ok(data) = entry.file_type() {
+                        if data.is_symlink() && walk_data.ignore_links {
+                            return None;
                         }
+                        if data.is_dir() || (walk_data.follow_links && data.is_symlink()) {
+                            return walk(entry.path(), permissions_flag, walk_data, depth + 1);
+                        }
+                        return build_node(
+                            entry.path(),
+                            vec![],
+                            walk_data.filter_regex,
+                            walk_data.invert_filter_regex,
+                            walk_data.use_apparent_size,
+                            data.is_symlink(),
+                            data.is_file(),
+                            walk_data.by_filecount,
+                            depth,
+                        );
                     }
                 } else {
                     permissions_flag.store(true, atomic::Ordering::Relaxed);

src/display.rs

@@ -88,7 +88,7 @@ impl DrawData<'_> {
         let mut num_not_my_bar = (chars_in_bar as i32) - num_bars as i32;
 
         let mut new_bar = "".to_string();
-        let idx = 5 - min(4, max(1, level));
+        let idx = 5 - level.clamp(1, 4);
 
         for c in self.percent_bar.chars() {
             num_not_my_bar -= 1;
@@ -141,10 +141,10 @@ pub fn draw_it(
     let longest_string_length =
         find_longest_dir_name(root_node, num_indent_chars, allowed_width, !use_full_path);
 
-    let max_bar_length = if no_percent_bars || longest_string_length + 7 >= allowed_width as usize {
+    let max_bar_length = if no_percent_bars || longest_string_length + 7 >= allowed_width {
         0
     } else {
-        allowed_width as usize - longest_string_length - 7
+        allowed_width - longest_string_length - 7
     };
 
     let first_size_bar = repeat(BLOCKS[0]).take(max_bar_length).collect();

src/filter.rs

@@ -8,6 +8,7 @@ use std::path::PathBuf;
 pub fn get_biggest(
     top_level_nodes: Vec<Node>,
     min_size: Option<usize>,
+    only_dir: bool,
     n: usize,
     depth: usize,
     using_a_filter: bool,
@@ -23,14 +24,14 @@ pub fn get_biggest(
     let mut allowed_nodes = HashSet::new();
 
     allowed_nodes.insert(root.name.as_path());
-    heap = add_children(using_a_filter, min_size, &root, depth, heap);
+    heap = add_children(using_a_filter, min_size, only_dir, &root, depth, heap);
 
     for _ in number_top_level_nodes..n {
         let line = heap.pop();
         match line {
             Some(line) => {
                 allowed_nodes.insert(line.name.as_path());
-                heap = add_children(using_a_filter, min_size, line, depth, heap);
+                heap = add_children(using_a_filter, min_size, only_dir, line, depth, heap);
             }
             None => break,
         }
@@ -41,15 +42,22 @@ pub fn get_biggest(
 fn add_children<'a>(
     using_a_filter: bool,
     min_size: Option<usize>,
+    only_dir: bool,
     file_or_folder: &'a Node,
     depth: usize,
     mut heap: BinaryHeap<&'a Node>,
 ) -> BinaryHeap<&'a Node> {
     if depth > file_or_folder.depth {
-        heap.extend(file_or_folder.children.iter().filter(|c| match min_size {
-            Some(ms) => c.size > ms as u64,
-            None => !using_a_filter || c.name.is_file() || c.size > 0,
-        }))
+        heap.extend(
+            file_or_folder
+                .children
+                .iter()
+                .filter(|c| match min_size {
+                    Some(ms) => c.size > ms as u64,
+                    None => !using_a_filter || c.name.is_file() || c.size > 0,
+                })
+                .filter(|c| if only_dir { c.name.is_dir() } else { true }),
+        )
     }
     heap
 }

src/main.rs

@@ -11,7 +11,9 @@ mod utils;
 
 use crate::cli::build_cli;
 use std::collections::HashSet;
+use std::io::BufRead;
 use std::process;
+use sysinfo::{System, SystemExt};
 
 use self::display::draw_it;
 use clap::Values;
@@ -19,6 +21,7 @@ use config::get_config;
 use dir_walker::{walk_it, WalkData};
 use filter::get_biggest;
 use filter_type::get_all_file_types;
+use rayon::ThreadPoolBuildError;
 use regex::Regex;
 use std::cmp::max;
 use std::path::PathBuf;
@@ -89,14 +92,28 @@ fn get_regex_value(maybe_value: Option<Values>) -> Vec<Regex> {
         .collect()
 }
 
+// Returns a list of lines from stdin or `None` if there's nothing to read
+fn get_lines_from_stdin() -> Option<Vec<String>> {
+    atty::isnt(atty::Stream::Stdin).then(|| {
+        std::io::stdin()
+            .lock()
+            .lines()
+            .collect::<Result<_, _>>()
+            .expect("Error reading from stdin")
+    })
+}
+
 fn main() {
     let options = build_cli().get_matches();
     let config = get_config();
+    let stdin_lines = get_lines_from_stdin();
 
-    let target_dirs = options
-        .values_of("inputs")
-        .expect("Should be a default value here")
-        .collect();
+    let target_dirs = match options.values_of("inputs") {
+        Some(values) => values.collect(),
+        None => stdin_lines.as_ref().map_or(vec!["."], |lines| {
+            lines.iter().map(String::as_str).collect()
+        }),
+    };
 
     let summarize_file_types = options.is_present("types");
 
@@ -135,6 +152,8 @@ fn main() {
 
     let by_filecount = options.is_present("by_filecount");
     let limit_filesystem = options.is_present("limit_filesystem");
+    let ignore_links = options.is_present("ignore_links");
+    let follow_links = options.is_present("dereference_links");
 
     let simplified_dirs = simplify_dir_names(target_dirs);
     let allowed_filesystems = limit_filesystem
@@ -150,15 +169,13 @@ fn main() {
         filter_regex: &filter_regexs,
         invert_filter_regex: &invert_filter_regexs,
         allowed_filesystems,
+        ignore_links,
+        follow_links,
         use_apparent_size: config.get_apparent_size(&options),
         by_filecount,
         ignore_hidden: config.get_ignore_hidden(&options),
     };
-    // Larger stack size to handle cases with lots of nested directories
-    rayon::ThreadPoolBuilder::new()
-        .stack_size(usize::pow(1024, 3))
-        .build_global()
-        .unwrap();
+    let _rayon = init_rayon();
 
     let iso = config.get_iso(&options);
     let (top_level_nodes, has_errors) = walk_it(simplified_dirs, walk_data);
@@ -168,6 +185,7 @@ fn main() {
         false => get_biggest(
             top_level_nodes,
             config.get_min_size(&options, iso),
+            config.get_only_dir(&options),
             number_of_lines,
             depth,
             options.values_of("filter").is_some() || options.value_of("invert_filter").is_some(),
@@ -191,3 +209,19 @@ fn main() {
         )
     }
 }
+
+fn init_rayon() -> Result<(), ThreadPoolBuildError> {
+    let large_stack = usize::pow(1024, 3);
+    let mut s = System::new();
+    s.refresh_memory();
+    let available = s.available_memory();
+
+    if available > large_stack.try_into().unwrap() {
+        // Larger stack size to handle cases with lots of nested directories
+        rayon::ThreadPoolBuilder::new()
+            .stack_size(large_stack)
+            .build_global()
+    } else {
+        Ok(())
+    }
+}